View
221
Download
1
Category
Preview:
Citation preview
DefinitionsCommon Algorithms
Applications
Minimum Spanning TreesAlgorithms and Applications
Varun Ganesan
18.304 Presentation
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Outline
1 DefinitionsGraph TerminologyMinimum Spanning Trees
2 Common AlgorithmsKruskal’s AlgorithmPrim’s Algorithm
3 Applications
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Outline
1 DefinitionsGraph TerminologyMinimum Spanning Trees
2 Common AlgorithmsKruskal’s AlgorithmPrim’s Algorithm
3 Applications
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Graphs
In graph theory, a graph is an ordered pair G = (V ,E)comprising a set of vertices or nodes together with a setof edges.
Edges are 2-element subsets of V which represent aconnection between two vertices.
Edges can either be directed or undirected.Edges can also have a weight attribute.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Graphs
In graph theory, a graph is an ordered pair G = (V ,E)comprising a set of vertices or nodes together with a setof edges.
Edges are 2-element subsets of V which represent aconnection between two vertices.
Edges can either be directed or undirected.Edges can also have a weight attribute.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Graphs
In graph theory, a graph is an ordered pair G = (V ,E)comprising a set of vertices or nodes together with a setof edges.
Edges are 2-element subsets of V which represent aconnection between two vertices.
Edges can either be directed or undirected.Edges can also have a weight attribute.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Graphs
In graph theory, a graph is an ordered pair G = (V ,E)comprising a set of vertices or nodes together with a setof edges.
Edges are 2-element subsets of V which represent aconnection between two vertices.
Edges can either be directed or undirected.Edges can also have a weight attribute.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Connectivity
A graph is connected when there is a path between everypair of vertices.
A cycle is a path that starts and ends with the samevertex.
A tree is a connected, acyclic graph.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Connectivity
A graph is connected when there is a path between everypair of vertices.
A cycle is a path that starts and ends with the samevertex.
A tree is a connected, acyclic graph.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Connectivity
A graph is connected when there is a path between everypair of vertices.
A cycle is a path that starts and ends with the samevertex.
A tree is a connected, acyclic graph.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Outline
1 DefinitionsGraph TerminologyMinimum Spanning Trees
2 Common AlgorithmsKruskal’s AlgorithmPrim’s Algorithm
3 Applications
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Spanning Trees
Formally, for a graph G = (V ,E), the spanning tree isE ′ ⊆ E such that:
∃u ∈ V : (u, v) ∈ E ′ ∨ (v ,u) ∈ E ′ ∀v ∈ VIn other words: the subset of edges spans all vertices.
Linear Graph
|E ′| = |V | − 1In other words: the number of edges is one less than thenumber of vertices, so that there are no cycles.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Spanning Trees
Formally, for a graph G = (V ,E), the spanning tree isE ′ ⊆ E such that:
∃u ∈ V : (u, v) ∈ E ′ ∨ (v ,u) ∈ E ′ ∀v ∈ VIn other words: the subset of edges spans all vertices.
Linear Graph
|E ′| = |V | − 1In other words: the number of edges is one less than thenumber of vertices, so that there are no cycles.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
Spanning Trees
Formally, for a graph G = (V ,E), the spanning tree isE ′ ⊆ E such that:
∃u ∈ V : (u, v) ∈ E ′ ∨ (v ,u) ∈ E ′ ∀v ∈ VIn other words: the subset of edges spans all vertices.
Linear Graph
|E ′| = |V | − 1In other words: the number of edges is one less than thenumber of vertices, so that there are no cycles.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Graph TerminologyMinimum Spanning Trees
What Makes A Spanning Tree The Minimum?
MST Criterion: When the sum of the edge weights in aspanning tree is the minimum over all spanning trees of a graph
Figure: Suppose (b, f ) is removed and (c, f ) is added...
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Outline
1 DefinitionsGraph TerminologyMinimum Spanning Trees
2 Common AlgorithmsKruskal’s AlgorithmPrim’s Algorithm
3 Applications
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with |V | disjoint components.
Consider lesser weight edges first to incrementally connectcomponents.
Make certain to avoid cycles.
Continue until spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with |V | disjoint components.
Consider lesser weight edges first to incrementally connectcomponents.
Make certain to avoid cycles.
Continue until spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with |V | disjoint components.
Consider lesser weight edges first to incrementally connectcomponents.
Make certain to avoid cycles.
Continue until spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with |V | disjoint components.
Consider lesser weight edges first to incrementally connectcomponents.
Make certain to avoid cycles.
Continue until spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Pseudocode
Kruskal’s Algorithm1 A = 02 foreach v ∈ G.V :3 MAKE-SET(v)4 foreach (u, v) ordered by weight(u, v), increasing:5 if FIND-SET (u) 6= FIND-SET (v ):6 A = A ∪ (u, v)7 UNION(u, v)8 return A
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Example
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Example
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Example
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Example
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Example
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Example
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Proof Of Correctness
Spanning Tree ValidityBy avoiding connecting two already connected vertices,output has no cycles.If G is connected, output must be connected.
MinimalityConsider a lesser total weight spanning tree with at leastone different edge e = (u, v).If e leads to less weight, then e would have beenconsidered before some edge that connects u and v in ouroutput.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Proof Of Correctness
Spanning Tree ValidityBy avoiding connecting two already connected vertices,output has no cycles.If G is connected, output must be connected.
MinimalityConsider a lesser total weight spanning tree with at leastone different edge e = (u, v).If e leads to less weight, then e would have beenconsidered before some edge that connects u and v in ouroutput.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Proof Of Correctness
Spanning Tree ValidityBy avoiding connecting two already connected vertices,output has no cycles.If G is connected, output must be connected.
MinimalityConsider a lesser total weight spanning tree with at leastone different edge e = (u, v).If e leads to less weight, then e would have beenconsidered before some edge that connects u and v in ouroutput.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Proof Of Correctness
Spanning Tree ValidityBy avoiding connecting two already connected vertices,output has no cycles.If G is connected, output must be connected.
MinimalityConsider a lesser total weight spanning tree with at leastone different edge e = (u, v).If e leads to less weight, then e would have beenconsidered before some edge that connects u and v in ouroutput.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Proof Of Correctness
Spanning Tree ValidityBy avoiding connecting two already connected vertices,output has no cycles.If G is connected, output must be connected.
MinimalityConsider a lesser total weight spanning tree with at leastone different edge e = (u, v).If e leads to less weight, then e would have beenconsidered before some edge that connects u and v in ouroutput.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Proof Of Correctness
Spanning Tree ValidityBy avoiding connecting two already connected vertices,output has no cycles.If G is connected, output must be connected.
MinimalityConsider a lesser total weight spanning tree with at leastone different edge e = (u, v).If e leads to less weight, then e would have beenconsidered before some edge that connects u and v in ouroutput.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Outline
1 DefinitionsGraph TerminologyMinimum Spanning Trees
2 Common AlgorithmsKruskal’s AlgorithmPrim’s Algorithm
3 Applications
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with one (any) vertex.
Branch outwards to grow your connected component.
Consider only edges that leave the connected component.
Add smallest considered edge to your connectedcomponent.
Continue until a spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with one (any) vertex.
Branch outwards to grow your connected component.
Consider only edges that leave the connected component.
Add smallest considered edge to your connectedcomponent.
Continue until a spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with one (any) vertex.
Branch outwards to grow your connected component.
Consider only edges that leave the connected component.
Add smallest considered edge to your connectedcomponent.
Continue until a spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with one (any) vertex.
Branch outwards to grow your connected component.
Consider only edges that leave the connected component.
Add smallest considered edge to your connectedcomponent.
Continue until a spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Main Idea
Start with one (any) vertex.
Branch outwards to grow your connected component.
Consider only edges that leave the connected component.
Add smallest considered edge to your connectedcomponent.
Continue until a spanning tree is created.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Example
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Kruskal’s AlgorithmPrim’s Algorithm
Proof of Correctness
Both the spanning tree and minimality argument are nearlyidentical for Prim’s as they are for Kruskal’s.
Varun Ganesan MSTs
DefinitionsCommon Algorithms
Applications
Some Applications
Taxonomy
Clustering Analysis
Traveling Salesman Problem Approximation
Varun Ganesan MSTs
Recommended